Germ expansion for 𝑺 𝑳 ( 𝟐 ) \bm{SL(2)} in arbitrary characteristics
Jean-Pierre Labesse · 2025
Read Germ expansion for 𝑺 𝑳 ( 𝟐 ) \bm{SL(2)} in arbitrary characteristics by Jean-Pierre Labesse (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-05003.txt.
Introduction
Jean-Pierre Labesse extends a classic expansion for orbital integrals to fields of any characteristic, in a 2025 paper. The setting is the group SL(2) over a local field. Shalika's germ expansion writes such integrals near the identity as a sum over unipotent classes. That sum makes sense only when the class set stays tame.
In characteristic two it does not: the classes swell to an uncountable crowd, and the classic expansion breaks. The endoscopic expansion, by contrast, always exists. Labesse shows it yields a germ expansion for elliptic elements in every characteristic. Where the field's trait is not two, the two forms match up to a Fourier twist. Where it is two, the endoscopic form is genuinely new. A conjecture for general reductive groups closes the paper, pointing past the rank-one case.
Catalog record
| CREATOR | Jean-Pierre Labesse |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 23 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-05003
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Germ expansion for 𝑺 𝑳 ( 𝟐 ) \bm{SL(2)} in arbitrary characteristics Jean-Pierre Labesse Abstract . Let F F be a local field of characteristic p p and G G be a connected reductive group over F F . Recall that Shalika’s germ expansion [ Sh ] of orbital integrals of regular semisimple elements near the identity, when it exists, is a sum indexed by the set 𝒰 G \mathcal{U}^{G} of unipotent conjugacy classes in G ( F ) G(F) . Observe that if G = S L ( 2 ) G=SL(2) the set 𝒰 G \mathcal{U}^{G} is always compact; it is finite if p ≠ 2 p\neq 2 while it is uncountable if p = 2 p=2 . As a consequence, Shalika’s germ expansion for elliptic elements does not make sense if p = 2 p=2 . On the other hand the endoscopic expansion of elliptic orbital integrals always exists and yields a germ expansion equivalent if p ≠ 2 p\neq 2 (up to a Fourier transform) to Shalika’s germ expansion but is new if p = 2 p=2 . 1. Introduction The asymptotic behaviour of orbital integrals of semisimple elements in a connected reductive group over F is controlled by Shalika’s germs when F is a local field of characteritic zero and more generally whenever the number of unipotent orbits is finite and no inseparability occurs. The simplest case where these hypothesis fail is SL(2,F) with F of characteristic 2 since, for such a group, there is an uncountable set of unipotent orbits. Shalika’s germ expansion is a very useful tool in local harmonic analysis and in particular one would need a substitute for them to extend, in positive characteristic, the proof by Kottwitz of Weil’s conjecture on Tamagawa numbers for global fields of characteristic zero. Observe that the case of quasi-split groups can readily be treated using Langlands-Lai technique since the spectral decomposition is available, thanks to Morris, in arbitrary characteristic. But inner forms are, for the time being out of reach, via Kottwitz’ techniques, in small characteristics. We understand there is a proof by completely different methods for positive characteristics but one would like to have a uniform approach independent of the characteristic. This is our main motivation for looking more carefully at a very simple special case. It turns out that a natural substitute valid in any characteristic can be given, for SL(2) , using endoscopy. This appeared, with detailed proofs, in the lecture notes [ L ] where local and global aspects of the contribution of unipotent elements to harmonic analysis for SL(2) in arbitrary characteristics are discussed. It seems reasonable to expect that similar techniques will provide, for arbitrary reductive groups, the substitue one is looking for. This will be stated as a conjecture. But the state of art, in (small with respect to the order of the Weyl group) positive characteristic, does not allow to go much farther for the time being. 2. Notation From now on, except in section 8 , let G=SL(2) and {\widetilde{G}}=GL(2) . We denote by Z the center of G and by {\widetilde{Z}} the center of {\widetilde{G}} . Then {\widetilde{G}} act by automorphisms, induced by conjugation, on G . Let F be a non-archimedean local field of charateristic p . The locally compact groups G(F) and {\widetilde{G}}(F) are endowed with Haar measures denoted dg and d{\widetilde{g}} respectively. Now let f\in\mathcal{C}_{c}^{\infty}(G(F)) and x\in G(F) . We denote by I_{x} the centralizer of x in G (as a scheme) and by {\widetilde{I}}_{x} its centralizer in {\widetilde{G}} . Observe that I_{x}(F) is always unimodular. This allows to define orbital integrals: \mathcal{O}_{G}(x,f)=\int_{I_{x}(F)\backslash G(F)}f(g^{-1}xg)d\dot{g}\quad\hbox{and}\quad\mathcal{O}_{\widetilde{G}}(x,f)=\int_{{\widetilde{I}}_{x}(F)\backslash{\widetilde{G}}(F)}f({\widetilde{g}}^{-1}x{\widetilde{g}})d\dot{\widetilde{g}} where d\dot{g} and d\dot{\widetilde{g}} are quotient measures. The quotient group Q_{F}:=(F^{\times})^{2}\backslash F^{\times} is always compact. It is finite if p\neq 2 but is uncountable when p=2 . The determinant induces an isomorphism {\widetilde{Z}}(F)G(F)\backslash{\widetilde{G}}(F)\simeq Q_{F}~. The set \mathcal{U}^{G} of conjugacy classes of non trivial unipotent elements in G(F) is easily seen to be also in (non canonical) bijection with Q_{F} . We shall denote by \mathcal{K} the Pontryagin dual of Q_{F} . Its elements are of order 2 and Class Field theory shows that they are in bijection with isomorphism classes of separable quadratic extensions E/F . The trivial character corresponds to the split extension E=F\oplus F . Let E/F be a separable quadratic extension (that may be split). We denote by e\mapsto\overline{e} the non trivial Galois automorphism (when E=F\oplus F it acts by the exchange of coordinates). Let E^{1} denote the subgroup of elements of norm 1 in E^{\times} . We denote by {\widetilde{T}}_{E/F} the restriction of scalar for E/F of \mathbb{G}_{m} and by T_{E/F} the subgroup of elements of norm 1. In particular we have {\widetilde{T}}_{E/F}(F)=E^{\times}\quad\hbox{and}\quad T_{E/F}(F)=E^{1}~. We denote by {\varepsilon_{E/F}} the character of F^{\times} associated to the quadratic extension E/F by Class Field theory. If E is a field {\varepsilon_{E/F}} is the non trivial character of N_{E/F}E^{\times}\backslash F^{\times} ; it is of order 2, while {\varepsilon_{E/F}} is the trivial character if E=F\oplus F . Let T be a (maximal) torus in G and {\widetilde{T}} its centralizer in {\widetilde{G}} . There is a separable quadratic extension E/F such that T (resp. {\widetilde{T}} ) is isomorphic to T_{E/F} (resp {\widetilde{T}}_{E/F} ). Conversely for any separable quadratic extension E/F there is a torus T\subset G such that T is isomorphic to T_{E/F} . One should be warned that, given E , there may be more than one conjugacy class of tori in G isomorphic to T_{E/F} . 3. Stable conjugacy Two elements in G(F) (resp. {\widetilde{G}}(F) ) will be said to be stably conjugate if they are conjugate in G(\overline{F}) (resp. {\widetilde{G}}(\overline{F}) ). Observe that, in the literature, stable conjugacy is only defined for semisimple elements and even for them it is slightly more subtle for non strongly regular elements in arbitrary reductive group (see for example [ MW , section I.4.4] ). But this simple definition works nicely for all elements in groups under consideration here. The set C(x) of conjugacy orbits under G(F) inside the set of rational points in the geometric orbit is parametrized by C(x)\simeq\mathrm{Coker}[H^{0}_{f}(F,G)\to H^{0}_{f}(F,I_{x}\backslash G)]\simeq\ker[H^{1}_{f}(F,I_{x})\to H^{1}_{f}(F,G)] where the H^{i}_{f} are flat cohomology sets. We have to use flat topology since, in characteristic 2, the scheme I_{x} has an infinitesimal factor when x=zu with z\in Z(F) and u a non trivial unipotent. For all other elements one could use étale cohomology as well since étale and flat cohomology coincide for smooth connected groups and one may forget about the index f . Since H^{1}_{f}(F,G) is trivial we have C(x)\simeq H^{1}_{f}(F,I_{x})~. Then standard cohomology results show that C(x) is trivial if x belongs to a split torus, it has two elements if x is regular semisimple in a non split torus and finally C(x)\simeq Q_{F} when x has a non-trivial unipotent factor in his Jordan decomposition. In particular it is an uncountable set when F is of characteristic 2. We also observe that if T is a torus in G H^{0}(F,T\backslash G)={\widetilde{T}}(F)\backslash{\widetilde{G}}(F)~. Lemma 3.1 . In G(F) (resp. {\widetilde{G}}(F) ) two elements are stably conjugate if and only if they are conjugate under {\widetilde{G}}(F) . Proof. Firstly one observes that since G(\overline{F}).{\widetilde{Z}}(\overline{F})={\widetilde{G}}(\overline{F}) conjugacy under G(\overline{F}) and {\widetilde{G}}(\overline{F}) coincide. Then it